Vol. 451: 75–92, 2012 MARINE ECOLOGY PROGRESS SERIES Published April 11 doi: 10.3354/meps09587 Mar Ecol Prog Ser OPENPEN ACCESSCCESS Both below- and aboveground shoalgrass structure influence whelk predation on hard clams Seiji Goshima1,*, Charles H. Peterson2 1Faculty of Fisheries Sciences, Hokkaido University, Hakodate, Hokkaido 041-8611, Japan 2Institute of Marine Sciences, University of North Carolina at Chapel Hill, Morehead City, North Carolina 28557, USA ABSTRACT: Seagrass introduces above- and belowground biogenic structure and complexity into otherwise relatively unstructured sand-flat habitat. Observation of burial depths by unmarked and ribbon-tagged knobbed whelks Busycon carica in sand and dense and sparse shoalgrass Halodule wrightii confirmed that shoalgrass structure inhibits whelk burrowing. Whelks showed a strong preference for sand-flat habitat during spring and summer, creating a partial refuge from whelk predation for hard clams Mercenaria mercenaria inside shoalgrass. By early autumn, shoal- grass blades sloughed off, reducing aboveground biomass by over 81%, leading to a massive invasion of seagrass habitat by whelks now exhibiting no habitat preference. The enhanced per- meability of the seagrass bed coincided with the large decline in aboveground shoalgrass bio- mass, a measure of strength of the physical boundary contrast (BC). As shown by tethering whelks in sand, intact shoalgrass, and shoalgrass with aboveground structure removed by clipping, belowground vegetation played a stronger role than aboveground vegetation in limiting the effi- ciency of predation on hard clams. From summer into autumn, average daily production of empty shells of whelk-consumed hard clams increased, with some evidence of greater increases within seagrass habitat, implying a seasonal breakdown of the refuge function of seagrass. The mecha- nism by which a strong BC inhibits whelk entry into and use of seagrass habitat is unresolved, per- haps acting as a physical barrier to passage into seagrass or alternatively acting as a behavioral cue indicative of expectation of lower predation efficiency associated with intact but not with clipped seagrass. KEY WORDS: Habitat structure · Boundary contrast · Busycon carica · Mercenaria mercenaria · Halodule wrightii · Permeability · Temporal refuge · Seagrass Resale or republication not permitted without written consent of the publisher INTRODUCTION between adjoining patches of different habitat types is variable, dependent not only on patch size and rel- Suitable habitats provide animals with a variety ative mobility of the occupants, but also on the per- and an abundance of acceptable food resources, liv- meability of the patch boundary, which can have ing spaces, refuges from predation and unfavorable ‘hard edges’ inhibiting passage or ‘soft edges’ allow- environmental stresses, and intrinsic advantages ing ready passage (Wiens et al. 1985, Stamps et al. over potential competitors. At some spatial scale, 1987). The resultant ‘boundary contrast’ (BC) is de - all such suitable habitats occur as discontinuous fined as the magnitude of difference in structural patches, often on relatively small spatial scales at habitat metrics across patch interfaces (Stamps et al. which a mosaic of patch types exists within appar- 1987, Holmquist 1998). The greater the BC, the less ently similar environments (Robbins & Bell 1994, permeable is the barrier created by this rapid struc- 2000, Vidondo et al. 1997). The degree of isolation tural transition at the patch boundary. *Email: [email protected] © Inter-Research 2012 · www.int-res.com 76 Mar Ecol Prog Ser 451: 75–92, 2012 The BC can influence not only rates of emigration invertebrates, and we conducted experiments to iso- and immigration connecting suitable patches of habi- late effects of aboveground from belowground struc- tat and the surrounding less hospitable areas but also ture on spatio-temporal dynamics of the benthic pre- predation pressure and environmental stress on indi- dation process. We hypothesized that the structural viduals that may tend to accumulate at relatively features of dense seagrass create a strong BC at the impassable habitat edges (strong BC). Predator−prey junction with a sand flat that inhibits permeability of encounter rates are often higher and environmental the seagrass edge to burrowing predators. Recogniz- conditions can be more severe along patch edges ing that burrowing predators that do penetrate the than in patch interiors, thereby reducing prey or boundary experience interference from the seagrass even predator survival (edge effect: Andren & Angel- structures, including suppressed digging ability stam 1988, Saunders et al. 1991, Andren 1992, Irlandi (Vince et al. 1976, Brenchley 1982, Peterson 1982, et al. 1995, Fagan et al. 1999). In cases of a weak BC, Heck & Wilson 1987, Irlandi 1997), we conducted potential predators and unfavorable environmental experimental and observational tests to quantify conditions may readily penetrate deeply into patches effects of below- versus aboveground shoalgrass of habitat otherwise favorable for prey species. In structure on behavior and efficiency of predation by contrast, a strong BC may serve as a barrier to per- a predatory benthic gastropod on its infaunal bivalve meability of physical environmental stressors and prey. We expect that the strong BC of a dense sea- predators, creating functional refuges inside such grass bed would also buffer environmental stresses, patches (Holmquist 1998). These boundary effects such as strong current flows, deep into seagrass consequently can modify spatial distributions and patches (Irlandi 1996, Peterson et al. 2004). population dynamics of patch inhabitants (Wiens Our focus on strength of the BC in exploring one 1995, Bender et al. 1998, Hovel & Regan 2008). possible mechanistic basis for understanding how During growth and development of seagrass habi- aboveground habitat structure may contribute to pro- tat patches, the plants spread via seed release, trans- viding a refuge from burrowing benthic predators is port, settlement, and germination as well as via vege - related to the widely acknowledged importance of tative propagation, typically resulting in a mosaic of habitat complexity. Functions of habitat complexity vegetated patches of varying sizes within a back- have been explored extensively in seagrass beds as a ground of unvegetated sediments (Thayer et al. 1984, model system amenable to careful observational and Duarte & Sand-Jensen 1990, Bowden et al. 2001). experimental hypothesis testing (e.g. Brenchley 1982, Seagrass patches and unvegetated sediments may Blundon & Kennedy 1982, Peterson 1982, Irlandi even reflect an example of multiple stable or at least 1997, Hovel & Regan 2008). Unfortunately, the cur - persistent states (Peterson 1984) because the higher rent view that belowground structure sufficiently current flows over unvegetated bottom inhibit sea- explains how seagrass habitat complexity affords in- grass colonization, and the seagrass beds act to slow faunal prey a substantial refuge from burrowing pre - currents and induce fine sediment and seed deposi- dators fails to explain many anomalous patterns. For tion, sustaining the seagrass plants and attendant example, although higher densities of hard clams muddier sediments in that location (Fonseca et al. Mercenaria mercenaria are typically observed in sea- 2007). grass than in surrounding sand flats (Peterson 1982, Seagrass beds provide 2 different types of structure 1986, Peterson & Beal 1989), some studies have re- that is absent from unvegetated sand and mud flats: vealed trivially small differences in density (Peterson aboveground shoots and leaves that provide struc- et al. 1984, Micheli 1997) or in age and size composi- ture in the water column, and roots and rhizomes that tion of the clams (Peterson et al. 1984), and at least bind the sediments below ground and enhance com- one study revealed lower clam densities in seagrass paction and bottom hardness (Peterson 1982). The than in nearby sand habitat (Nakaoka 2000). Further- aboveground structure provides refuge for nektonic more, the shells of hard clams recently consumed by species from visually orienting predators like fishes, whelks (as determined from characteristic shell dam- and belowground structure provides infaunal inver- age: Peterson 1982) are commonly observed inside tebrates a refuge from burrowing predators like seagrass beds in late summer (S. Goshima pers. obs.). crabs and whelks (Peterson 1982, see reviews by These observations, apparently inconsistent with the Orth et al. 1984, Orth 1992). For this new study, we structural refuge hypothesis, may conceivably be rec- focused novel attention on how the aboveground onciled if seasonal changes in aboveground seagrass seagrass structure, which can change seasonally, structure substantially alter the BC and allow whelks may also affect the behavior of predatory benthic to penetrate the seagrass edge and prey more readily Goshima & Peterson: Habitat boundary contrasts affect predators 77 on the hard clams inside. Seagrasses are known to 0.1−0.3 m at low tides to 1.1−1.3 m at spring high shed their leaves in late summer and autumn in North tides. Salinities exceed 34 psu in summer and Carolina (USA), when water temperatures reach high autumn and 32 psu in winter and spring in Bogue levels (Thayer et al. 1984), raising the possibility that Sound, except during occasional heavy rainstorms, the predation refuge is temporally